paper-with-me

Papers

Information Gain-Guided Causal Intervention for Autonomous Debiasing Large Language Models

2025-04-17 · Zhouhao Sun, Xiao Ding, Li Du, Yunpeng Xu, Yixuan Ma, Yang Zhao, Bing Qin, Ting Liu

Despite significant progress, recent studies indicate that current large language models (LLMs) may still capture dataset biases and utilize them during inference, leading to the poor generalizability of LLMs. However, due to the diversity of dataset biases and the insufficient nature of bias suppression based on in-context learning, the effectiveness of previous prior knowledge-based debiasing methods and in-context learning based automatic debiasing methods is limited. To address these challenges, we explore the combination of causal mechanisms with information theory and propose an information gain-guided causal intervention debiasing (IGCIDB) framework. This framework first utilizes an information gain-guided causal intervention method to automatically and autonomously balance the distribution of instruction-tuning dataset. Subsequently, it employs a standard supervised fine-tuning process to train LLMs on the debiased dataset. Experimental results show that IGCIDB can effectively debias LLM to improve its generalizability across different tasks.

📄 PDF Abstract BibTeX arXiv:2504.12898

Code (0)

등록된 구현이 없습니다.

Tasks

DiversityIn-Context Learning

Similar Papers 제목 키워드 기반

A Generalizable Physics-guided Causal Model for Trajectory Prediction in Autonomous Driving

2026-02-15 · Zhenyu Zong, Yuchen Wang, Haohong Lin, Lu Gan 외 arxiv

Trajectory prediction for traffic agents is critical for safe autonomous driving. However, achieving effective zero-shot generalization in previously unseen domains remains a significant challenge. Motivated by the consi…

Zero-shot GeneralizationTrajectory PredictionAutonomous Driving

CRRL: A Causality-Based Reinforcement Learning Framework for Autonomous System Recovery

2026-07-03 · Safia Fatima, Kai Olav Ellefsen, Leon Moonen arxiv

Traditional reinforcement learning (RL) for recovery in autonomous systems lacks causal understanding and generalizes poorly to novel failure scenarios. RL policies often stall in failure states, spending up to 70% of an…

Reinforcement Learning

Probability trees and the value of a single intervention

2022-05-18 · Tue Herlau

The most fundamental problem in statistical causality is determining causal relationships from limited data. Probability trees, which combine prior causal structures with Bayesian updates, have been suggested as a possib…

Active Learning

Active Causal Experimentalist (ACE): Learning Intervention Strategies via Direct Preference Optimization

2026-02-02 · Patrick Cooper, Alvaro Velasquez arxiv

Discovering causal relationships requires controlled experiments, but experimentalists face a sequential decision problem: each intervention reveals information that should inform what to try next. Traditional approaches…

Domain Adaptation

Active learning of causal probability trees

2022-05-17 · Tue Herlau

The past two decades have seen a growing interest in combining causal information, commonly represented using causal graphs, with machine learning models. Probability trees provide a simple yet powerful alternative repre…

Active Learning